Atlas

Compare Models

Claude Opus 5.5 vs GPT-6 Astra

Claude Opus 5.5 leads by 0.5 capability points (≈95% interval −2.2 to +3.2): statistically tied · P(higher) ≈ 64%

Wins 177 of 341 shared benchmarks (133 lost, 31 tied)

Capability vs. relative cost
Claude Opus 5.5GPT-6 AstraClosed weightsOpen weights

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

Claude Opus 5.5vsGPT-6 Astracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins177 – 133 · 341 shared, 31 tied

Comparing matching setups only (201 benchmarks excluded) · suites counted once
← Claude Opus 5.5GPT-6 Astra →

Claude Opus 5.5vsGPT-6 Astrawins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 99 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6 Astra minus Claude Opus 5.5.

Clipping Bench v1.0Games20189514+80.6
VCT-Refusal (CAIS)Safety90.417.6−72.8
AA-PersonalityBench Informativeness indexChat & Writing90.529.6−61.5
Mercor enterprise risk (implicit-risk unsafe rate)Safety22.280.6−58.3
LisanBench Path LengthGames240297393−51.7
HPCT Refusal (CAIS)Safety100.050.0−50.0
Clocktower BenchGames15471892+49.5
Olam Social Poker EloGames15061550+47.8
TxBench Oligo Discovery V1 Clean RerunsScience1.449.2+47.8
Design Creativity BenchCode51.065.5+45.9
TxBench Antibody Discovery V1 Clean RerunsScience6.848.9+42.1
Short Form Bench v1.0Games23892803+41.0
CheatBenchSafety7.245.0−37.8
BioSecBench-Refusal V1 Clean RerunsSafety3.840.8+37.0
ActiveVisionMultimodal41.276.5+35.3
TxBench Preclinical Pharmacology V1 Clean RerunsScience26.359.3+33.0
AA-PersonalityBench Expressiveness indexChat & Writing90.558.1−32.7
MirrorCode (ML, +Private, 2L)Code78.946.7−32.2
EnterpriseBench: CoreCraftProfessional Work74.442.6−31.8
Medical Long Context Reasoning (MLCR-AA)Health & Medicine66.735.0−31.7
TasteVal (compute multiplier)Science2.31.6−30.8
LisanBench Difficulty-WeightedGames59633113−30.5
Concept Synth INDUCTION Challenge100 Round 1General QA65.094.0+29.0
SpireClimbGames5.834.0+28.2
SWE-bench Multimodal (Vals v1, 480 Tasks)Code70.243.3−26.9
Concept Synth INDUCTION Challenge64 Round 1General QA71.998.4+26.6
Humanity's Last Exam RMS Calibration Error (CAIS)General QA13.839.8−26.0
AA-PersonalityBench Conviction indexChat & Writing69.544.8−25.0
BioSecBench-Function V1 Clean RerunsScience73.648.6−25.0
ChessBench EloGames13491755+24.2
GlobeBenchGeneral QA26.349.6+23.2
HANDBOOK.md Agents v1.1Professional Work52.331.4−20.9
RedlineBenchLegal42.062.8+20.8
ScBench-Long V1 Clean RerunsScience6.827.3+20.5
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.020.0+20.0
HoneyBench v0.2 (hack rate)Safety31.650.8−19.1
SpatialBench-Long V1 Clean RerunsScience20.239.3+19.1
APEX-SWECode67.648.6−19.0
BioTIER RefusalSafety98.779.9−18.8
DataCenterBench Tier 1Agents21.740.0+18.3
ProgramBench - almost solved rollout percentageCode70.552.5−18.0
AA-PersonalityBench Imagination indexChat & Writing94.377.2−17.3
Terminal-Bench Science 0.1 (Vals)Science48.665.7+17.1
TextQuests Average Cumulative HarmSafety24.229.2−17.0
ChessBench CoherenceGames80.197.0+16.9
TextQuests Game ProgressGames57.374.2+16.9
LAB-Bench LitQA2 (SecureBio)Science71.154.8−16.3
DAYJOB: Data ScienceProfessional Work37.321.3−16.0
ALE-BenchCode27483164+14.7
sudo L7 (October 2026 rescoring)Code45.631.1−14.5
Search ArenaSearch12131231+14.0
BioResearchBenchScience18.04.0−14.0
LLM Debate BenchmarkChat & Writing17011606−13.6
ArXivMath 05/2026Math81.795.0+13.3
DAYJOB: HealthcareHealth & Medicine24.711.6−13.1
Mystery Game PuzzlesGames71.084.0+13.0
AA-Briefcase v1.1 EloAgents18011569−12.7
Ivo-micro1 ContractBenchLegal58.145.6−12.5
GDPval-AA v2.1Professional Work18661587−12.3
LABBench2Science73.185.4+12.3
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science25.413.3−12.1
ProgramBench (Mercor 100-task subset)Code20.08.0−12.0
EnigmaEvalGeneral QA57.869.7+11.9
SM BenchSafety81.869.9−11.9
DAYJOB: LawLegal25.136.9+11.8
ViBench 2.0Code48.560.3+11.8
Harvey HLAB — Criteria Pass RateLegal75.887.5+11.7
GDPval Extended (Mercor held-out rubric score)Professional Work75.864.1−11.7
Agent Initiative BenchmarkAgents7.89.0+11.3
τ³-Banking (Fireworks, banking_knowledge)Agents56.044.8−11.2
DeepsecBench — Revalidated V1Code26.737.8+11.1
ArXivMath 06/2026Math83.394.4+11.1
DeepsecBench — Revalidated V1 (233 issues)Code26.637.6+11.1
WellnessBenchSafety53.064.0+11.0
Agent Red Teaming k=100 (CAIS)Safety24.413.5+10.9
EpiBench V1 Clean RerunsScience39.028.3−10.7
Humanity's Last Exam (with tools)General QA67.757.2−10.5
SnorkelLegalLegal31.420.9−10.5
SciCode 1.0.1 (Artificial Analysis)Code66.956.5−10.4
MedXpertQA MM Extended (Mercor held-out)Health & Medicine71.260.8−10.4
MetagenomicsBench V0Science39.349.7+10.4
FACTS Benchmark SuiteGeneral QA65.375.7+10.4
ITBench-AAAgents38.248.6+10.4
MetagenomicsBench V1Science39.349.7+10.3
Qodo AI Code Review Benchmark (125 PRs)Code56.846.6−10.2
HardGeoBench <=100miMultimodal61.351.3−10.0
ZeroBench Main Questions pass^5Multimodal25.035.0+10.0
VIALSMultimodal40.249.9+9.7
AA-PersonalityBench Warmth indexChat & Writing84.875.3−9.6
AA-PersonalityBench Organization indexChat & Writing79.169.5−9.6
Extended NYT Connections 940 (Quadratic v1)Games88.598.1+9.6
LMArena Agent Arena Net Improvement (Steering Burden)Agents19.514.5−9.4
SpeciEval (Refusal-Aware Retries, Historical)Safety87.296.5+9.3
SpeciEval (Pooled English Runs)Safety87.296.5+9.3
JobBench - release-reported mean task rewardProfessional Work75.466.1−9.3
CVDP (517-problem set)Code64.154.9−9.2
Humanity's Last Exam DiamondGeneral QA54.663.8+9.2
IntPhys 2Multimodal68.377.4+9.2
MASKSafety89.498.5+9.1
Humanity's Sixth SenseMultimodal44.653.6+9.0
MMMU-Pro Extended (Mercor held-out)Multimodal51.060.0+9.0
ZeroBench (Main questions, pass@5)Multimodal43.052.0+9.0
Code2Video-AA EloCode15361427−8.8
Financial Coaching BenchFinance & Accounting76.267.5−8.6
ScBench V1 Clean RerunsScience55.463.9+8.5
AHK-EvalCode91.7100.0+8.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.046.7−8.3
PCT Political Inconsistency (CAIS)Safety29.137.4−8.3
ZeroBench Main Mean pass@1 (five samples)Multimodal35.043.2+8.2
SWE-TogetherCode68.860.6−8.2
GraphWalks BFS: Up to 128KGeneral QA90.698.7+8.1
WCB (SecureBio)Science54.562.6+8.1
BioSecBench-Structure V1 (58 evaluations)Science29.321.3−8.0
Eval Connections One-Shot with TrapsGames86.094.0+8.0
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine29.321.3−8.0
SWE Atlas v1.1 - OverallCode77.469.4−8.0
Toloka Arena (v3)Agents75.067.0−8.0
o11y-bench 1.1 Pass^3Agents93.885.9−7.8
ComplexConstraintsChat & Writing49.957.7+7.8
HackerRank-BenchCode51.243.5−7.7
SciCode (main problems)Science27.319.6−7.7
Decision-Bench v1.1Professional Work59.151.5−7.7
Gray Swan IPI Attack Success (k=15)Safety1.08.5−7.5
HardGeoBench Country AccuracyMultimodal67.560.0−7.5
GDP.xlsxProfessional Work30.322.9−7.4
Into the Unknown Integrity Node Pass RateScience90.497.8+7.4
HardGeoBench Avg DistanceMultimodal669908−7.4
HealthBenchHealth & Medicine64.356.9−7.4
BioSecBench-Surveillance V1 Clean RerunsScience55.848.5−7.3
ERQAMultimodal75.182.3+7.2
Snorkel Agentic Coding 2.0Code40.447.6+7.2
Terminal-Bench-Science 0.1Science61.954.8−7.1
Quant-Arena (iterative total score)Finance & Accounting44.351.3+7.0
SnorkelFinance 2.0Finance & Accounting18.611.6−7.0
Chess Puzzles (task version 2.0)Games67.074.0+7.0
LegalReasonBenchLegal46.539.5−7.0
Terminal-Bench 4.0Code64.857.9−6.9
ACCoRDGeneral QA88.481.7−6.7
PubMedQA (SecureBio)Health & Medicine87.981.2−6.7
Epiq LAER CorporateBenchProfessional Work75.482.0+6.6
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting48.254.6+6.4
Terminal-Bench 2.1Code83.589.9+6.4
Creative Writing v3 EloChat & Writing20512177+6.4
Legacy-Bench (Fireworks, avg@3)Code62.055.7−6.3
HiL-Bench (Human-in-Loop Benchmark)Agents59.065.3+6.3
OSWorld 2.1 offline - reported partial reward, step limit unspecifiedAgents87.280.9−6.3
FrontierCUA pass@1Agents36.042.2+6.2
APEX-Agents 1.1Professional Work81.375.1−6.2
ChessBench AccuracyGames79.385.0+6.1
BullshitBench v2 LeaderboardSafety63.069.0+6.0
GeneBench-Pro Public Release SubsetScience30.036.0+6.0
OSWorld 2.1 full - reported partial reward, step limit unspecifiedAgents81.875.8−6.0
MacroscopeBenchCode83.978.0−5.9
BrowseComp (Mercor 130-question subset)Search88.594.2+5.7
LMArena Text Arena EloChat & Writing15061474−5.6
Mercor enterprise risk (explicit-risk unsafe rate)Safety5.60.0+5.6
Conceptual Reasoning IndexIndexes77.973.9−5.5
GBENCH One-shot Coding GScoreGames73.979.2+5.3
micro1 PharmaBenchHealth & Medicine10.816.1+5.3
WMDP-Bio (SecureBio)Science96.190.8−5.3
SpatialBench V1 Clean RerunsScience72.867.5−5.2
AA-AnalystAgent (pass^5)Professional Work56.351.3−5.0
CloudOpsBench v1.0Agents47.842.8−5.0
GraphWalks BFS: 256K to 1M AggregateGeneral QA66.871.8+5.0
HardGeoBench <=10miMultimodal42.537.5−5.0
PostTrainBench Average (v1.1)Indexes49.344.3−5.0
AA-3D v1.0 EloCode12021168−4.9
Westworld Finance Diligence BenchFinance & Accounting59.454.5−4.9
MentalHealthBenchSafety52.457.3+4.9
SimpleBenchGeneral QA88.483.6−4.8
MBCT (SecureBio)Science68.763.9−4.8
EBR-bench v4 (card ban, single agent)Games71.476.2+4.8
LMArena Agent Arena Bash RecoveryAgents15.05.5−4.7
BuseyBench SVGMultimodal6.87.3+4.7
ChartographyMultimodal66.371.0+4.7
LAB-Bench ProtocolQA (SecureBio)Science94.089.3−4.7
ChessBenchGames65.370.0+4.7
AA-LCR v1.1Search84.780.0−4.7
Arena Alignment IndexSafety83.287.8+4.6
Bio Capabilities Index (SecureBio)Indexes171164−4.5
MaxxBenchCode76.471.9−4.5
Immigration BenchLegal72.167.7−4.4
Humanity's Last Exam (Text-Only)General QA57.553.1−4.4
DAYJOB: TaxFinance & Accounting9.24.8−4.4
DAYJOB: AccountingFinance & Accounting34.430.0−4.4
LMArena Agent Arena Praise vs ComplaintAgents40.331.7−4.3
Excel Modeling BenchmarkFinance & Accounting75.971.7−4.2
Snorkel SWE-bench CLICode16.612.4−4.2
FrontierCode 2.0 (Artificial Analysis)Code46.442.4−4.1
Agents' Last Exam (ALE)Agents38.234.2−4.0
BrokenArXiv 05/2026Math95.091.0−4.0
Gym-Anything-RT (real-time, live)Games5.09.0+4.0
VariantBench V1 Clean RerunsScience50.646.6−4.0
APEX-AccountingFinance & Accounting61.857.9−3.9
Learn2Play BenchGames80.176.2−3.9
LMCAGeneral QA68.264.4−3.9
AA-PersonalityBench Slop indexChat & Writing20.016.2−3.8
AA-PersonalityBench Transparency indexChat & Writing90.594.3+3.8
LibraryDesignBenchCode48.945.1−3.8
LVBenchMultimodal83.787.5+3.8
LMArena Agent Arena Bash ReliabilityAgents15.17.8−3.6
Into the Unknown Direction Node Pass RateScience80.083.6+3.6
GDP.pdfProfessional Work30.634.2+3.6
Machiavelli Deception (CAIS)Safety77.576.3+3.5
MedScribeHealth & Medicine91.487.9−3.5
APEX-1 (Mercor, mean score)Professional Work65.869.3+3.5
GPQA-Bio (SecureBio)Science90.987.4−3.5
SpatialViz-BenchMultimodal89.793.1+3.4
SimpleQA Verified (Epoch anti-abstention)General QA72.275.6+3.4
Furniture Assembly BenchmarkMultimodal83.380.0−3.3
LMArena Agent Arena Confirmed SuccessAgents20.714.0−3.3
DeepResearch Bench II (Mercor 130-task set)Search53.350.1−3.2
Excel Modeling Benchmark v1.1Finance & Accounting75.872.7−3.2
FrontierSWE V2 Worst@5Code52.055.1+3.1
LLM Creative Story WritingChat & Writing3.73.4−3.1
Hermes BenchAgents76.773.7−3.1
CrossSearch (pass^3)Search57.060.0+3.0
VCT (SecureBio)Science61.458.5−2.9
Mercor enterprise risk (benign-control completion rate)Safety100.097.2−2.8
OpenWorldEngineer (Mean@8)Code41.644.2+2.6
HieroglyphBenchMultimodal37.440.0+2.6
FrontierMath Tier 4 (v2) - PrivateMath95.097.6+2.6
VCT v2 (SecureBio)Science60.558.0−2.5
Dual-Arm Sim (task completion)Agents0.02.5+2.5
Senior SWE-BenchCode39.336.8−2.5
FrontierMath Tiers 1-3 (v2) - PrivateMath91.293.7+2.5
GDPxProfessional Work12.09.6−2.4
Ruby LLM Benchmarks — Program FixerCode77.575.1−2.4
RiemannBenchMath69.672.0+2.4
DAYJOB: FinanceFinance & Accounting23.921.5−2.4
Political Science LLM BenchmarkProfessional Work70.267.9−2.3
Harvey LAB: In-House Contracting (held-out)Legal3.05.3+2.3
Hebbia Financial Services BenchmarkFinance & Accounting74.672.3−2.3
Drone-BenchAgents97.395.1−2.2
Visa Agentic Commerce BenchmarkAgents62.264.4+2.2
ClockBench AccuracyMultimodal77.275.0−2.2
HLAB-Vals-1.1Legal9.27.1−2.1
SAB ResearchFinance & Accounting97.295.2−2.1
TaxCalcBench TY25Finance & Accounting42.044.0+2.0
MMLU ProGeneral QA93.491.5−1.9
LMArena Agent Arena Steering BurdenAgents18.614.8−1.9
Ramp Accounting BenchFinance & Accounting50.548.7−1.8
Longform Creative WritingChat & Writing84.682.8−1.8
HealthBench ProfessionalHealth & Medicine71.469.7−1.7
MissionReportProfessional Work63.561.8−1.7
ARC-AGI-2General QA93.395.0+1.7
DTBenchGeneral QA98.997.3−1.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.89.2−1.6
SnorkelRevOpsProfessional Work12.510.9−1.6
GDPval (Mercor 213-task rubric score)Professional Work88.386.7−1.6
LLM Chess EloGames15751614+1.6
Blueprint-Bench 2Multimodal0.50.5−1.5
SnorkelUnderwrite 2.0Finance & Accounting22.721.2−1.5
FrontierSWE V2 Best@5Code71.573.0+1.5
CritPt v2 (agentic, avg@5)Science81.783.1+1.4
MMLU-Bio (SecureBio)Science92.791.3−1.4
Roboflow Vision Evals (October 2026, five tasks)Multimodal82.083.4+1.4
BioSimBench 1.0Science14.515.9+1.4
AA-Omniscience IndexIndexes46.443.7−1.3
MedCodeHealth & Medicine49.848.5−1.3
DistressBenchSafety75.674.3−1.3
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine57.058.2+1.2
WARC-Bench (192-task test split)Agents90.389.1−1.2
Investor's Last ExamFinance & Accounting66.968.1+1.2
Vectara Hallucination LeaderboardSafety9.98.7+1.2
AutomationBench 1.0.6 (Zapier)Agents42.541.4−1.1
SWE Atlas - Codebase QnACode48.947.8−1.1
MedXpertQA MM (200-question set)Health & Medicine85.786.8+1.1
Roboflow Vision Evals (September 2026 protocol)Multimodal85.586.6+1.1
DAXBench v2Code95.796.8+1.1
FORTRESS Average Risk Score (ARS)Safety13.212.1+1.0
DualEntry Accounting AI BenchmarkFinance & Accounting82.281.2−1.0
OpenDS-WorldSearch38.039.0+1.0
PLE (Physical-Engineering's Last Exam)Professional Work49.348.3−1.0
ProofBench v1.1Math100.099.0−1.0
BrokenArXiv 06/2026Math100.099.1−0.9
AutoResearchExam (hidden-test AUARC)Agents0.60.6−0.9
SAB RankingForecasting-0.1-0.0+0.9
VoxelBench Text RatingGames26132625+0.9
Nimble Web ContextSearch71.872.6+0.8
Time Horizon Index: KSPAgents91.390.5−0.8
AdvancedIFChat & Writing82.882.0−0.8
Professional Reasoning Benchmark - LegalLegal47.648.4+0.8
FrontierCode 1.1 Extended - ScoreCode65.364.5−0.8
Prophet Arena Arena Score (Fixed Context)Forecasting85.584.7−0.8
SUPERChemScience79.680.3+0.7
ForecastBench Preliminary Dataset Brier IndexForecasting62.262.9+0.7
Physical-Engineering's Last Exam (hidden test)Professional Work46.445.8−0.6
Snorkel Digital Work IndexIndexes23.623.0−0.6
SAGEMath45.846.4+0.5
Remote Labor Index (RLI)Professional Work21.320.8−0.4
GameBench 2Games94.894.3−0.4
SnorkelManufacturingProfessional Work10.410.0−0.4
AutomationBench-AA - release-reported strict pass rateAgents41.241.6+0.4
CharXiv Extended (Mercor held-out)Multimodal84.785.1+0.4
Epoch Capabilities Index (ECI)Indexes169169+0.4
LiveBench 2.0Agents86.586.1−0.4
MindCube TinyMultimodal93.192.7−0.4
CharXiv (descriptive and reasoning combined)Multimodal96.496.1−0.3
BenchCAD - geometric overlapMultimodal96.295.9−0.3
Creative Writing v3 Rubric ScoreChat & Writing16.015.9−0.2
MMMU ProMultimodal86.686.4−0.2
LMArena Agent Arena Tool HallucinationAgents0.71.0+0.2
DeepSWE v1.1Code74.274.1−0.1
ATLAS-Finance v1.0Finance & Accounting11.411.3−0.1
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.286.20.0
FrontierMath ErdősMath2.92.90.0
AIME 2025Math100.0100.00.0
ARC-AGI-1 Semi-PrivateGeneral QA98.598.50.0
CadQueryEval (Corrected Geometry)Code100.0100.00.0
CritPt (avg@5)Science31.731.70.0
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Arbitration International Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Banking FinanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Bankruptcy RestructuringLegal0.00.00.0
Harvey's Legal Agent Benchmark - Capital MarketsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal0.00.00.0
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.00.00.0
Harvey's Legal Agent Benchmark - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Environmental ESGLegal0.00.00.0
Harvey's Legal Agent Benchmark - Funds Asset ManagementLegal0.00.00.0
Harvey's Legal Agent Benchmark - ImmigrationLegal0.00.00.0
Harvey's Legal Agent Benchmark - InsuranceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Intellectual PropertyLegal0.00.00.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Litigation Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Real EstateLegal0.00.00.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.00.00.0
Harvey's Legal Agent Benchmark - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0
Hedge-BenchFinance & Accounting44.244.20.0
Into the UnknownScience0.00.00.0
Next.js AI Agent EvaluationsAgents100.0100.00.0
ObviousBenchGeneral QA100.0100.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0
Snorkel WorkplaceAgentsProfessional Work14.414.40.0
Terminal-Bench 4.0 (Artificial Analysis)Code59.659.60.0

Claude Opus 5.5vsGPT-6 Astranormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 99 othersVillage LLM Atlasatlas.village.dev

Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…